The Prevalence of Chronic Kidney Disease in Hypertensive Patients in Primary Care in Hong Kong: A Cross-Sectional Study

Background To identify the prevalence of Chronic Kidney Disease (CKD) in Chinese hypertensive population managed in a local and to explore its associated risk factors. using Collaboration 2012 used analyze continuous variables and the Chi-squared test used for categorical Multivariate Logistic regression was used to examine the association between CKD and variable associated All statistical tests two-sided, and risk factors for CKD among HT patients.


Introduction
Chronic Kidney Disease (CKD) is a worldwide public health problem [1]. In Kidney Disease Improving Global Outcome (KDIGO) 2012 clinical guideline [2], CKD is defined as abnormalities of kidney structure or function, present for more than 3 months, with implications for health. It is confirmed to be associated with an increased risk of cardiovascular comorbidities and mortality [3,4] as well as progression to dialysis dependent End Stage Renal Disease (ESRD) [5,6] . In our daily practice, CKD refers to CKD stage 3 to 5 in the KDIGO CKD staging system, which is defined as Estimated Glomerular Filtration Rate (eGFR) being less than 60 ml/min/1.73m 2 . Extensive studies have shown that this group of patients carries a particularly high risk for complications and adverse outcomes [7,8] . The prevalence of CKD in the general population varies in regions, e.g. 8.7% in selected countries in Africa, 13.1% in Indian subcontinent, 14.7% in Australia, 15.5% in North America, 18.4% in Europe, 13.7% in Japan and South Korea, 13.2% in Greater China region, with considerable international variation [9]. In Hong Kong, a screening study showed the prevalence of positive (≥1+) urine dipstick for protein, glucose, blood, protein or blood, any urine abnormality was 3.2%, 1.7%, 13.8%, 16%, 17.4%, respectively in apparently "healthy" (asymptomatic and without history of DM, HT, or CKD) individuals [10]. Hypertension (HT) is a well-recognized 2. Method

Aim
The objective of this study is to explore the prevalence of CKD in hypertensive patients in a public primary care clinic and to identify the possible associated factors.

Study design and setting
It is a cross-sectional study in a public primary care clinic. Yau Ma Tei General Outpatient Clinic is one of three public primary care clinics mainly serving but not restricted to an urban district with population around 330000 (Census and Statistics Department Hong Kong Special Administrative Region 2019).

Inclusion criteria
Chinese HT patients with International Classification of Primary Care (ICPC) code K86 (uncomplicated HT) or K87 (complicated HT) in the Clinical Management System (CMS), who had at least one follow up in a public primary care clinic from 01/01/2018 to 30/06/2018 and had at least two sets of serum renal function tests (RFT) done 3 months apart in the previous 3 years were included. (females) or 0.9 (males); α = -0.329 (females) or -0.411 (males); min = indicates the minimum of SCr/κ or 1; max = indicates the maximum of SCr/κ or 1; age = years; Scr in mg/dl.

Persistence of kidney abnormality
The latest two serum creatinine levels were retrieved from the CMS, which were at least 3 months apart. The mean eGFR was used for diagnosis and staging of CKD.

Sample size estimation
One proportion cross-sectional formula was used to calculate the sample size (website http://www2.ccrb.cuhk.edu.hk/stat/epistud.htm). Assume Probability of type 1 error is 0.01, prevalence proportion p is 0.15, estimated effect size is 1, desired level of absolute precision is 0.03, the required sample size is 940. To allow the room for sample exclusion (~20%), a total of 1200 patients was randomly selected by online (https://www.randomizer.org/) generated random numbers for data analysis. Briefly, all the included patients were listed in order of their outpatient case numbers, then a list of random numbers was generated from the research randomizer from which the 1200 patients to be included were selected.

Statistical analysis
Statistical calculations were completed using SPSS 19 (IBM SPSS Statistics version 19). Continuous variables were described as mean and standard deviation, while qualitative variables were expressed as numbers and percentage. Ttest was used to compare quantitative variables and the Chi-squared test for categorical variables. Mantel-Haenszel test was used for trend between age groups and CKD prevalence. Multivariate logistic regression analysis was used to identify the risk factors for the presence of CKD. All statistical tests were two-sided, and a P value of less than 0.05 was considered significant.

Study population and sampling process
From 01/01/2018 to 30/06/2018, totally 17,698 HT patients had at least one follow-up visit in the clinic. Among them, 1200 patients were randomly selected, from which 207 cases were excluded, including 101 Non-Chinese, 2 wrongly labeled HT patients and 104 cases who had no repeated RFT tests 3 months apart. Therefore, the remaining 993 cases were included in the final analysis. The selection and sampling process was summarized in Figure 1. The demographics and comorbidities of HT patients were demonstrated in Table 1. Among the 993 patients included in data analysis, 489 were female and 504 were male, with an average age of 68.9±10.9 years. The smoking status, BMI and comorbidities were retrieved.

Prevalence of CKD and distribution in age groups
As shown in Table 2, as defined by eGFR < 60 ml/min/1.73m 2 , the prevalence of CKD was 17.5% in male, 13 Figure 2 showed the prevalence of CKD in various age groups with apparent increasing trend of prevalence of CKD as age increased, the elder the age group, the higher the prevalence of CKD (trend p<0.001).

Figure 2:
Prevalence of CKD by age groups of HT patients (Trend P< 0.001)

Discussion
To the best of our knowledge, this is the first study to describe the CKD prevalence among unselected HT cases managed in the primary care setting in Hong Kong. We found that defined by eGFR < 60 ml/min/1.73m 2 , the prevalence of CKD was 17.5% in male, 13.1% in female, and 15.3% overall. Older age, history of CHF, DM, gout, lower level of HDL and presence of proteinuria or albuminuria are the predisposing factor of existence of CKD.
Renal function test is essential to the diagnosis and staging of CKD. However, serum creatinine alone is not reliable to assess the renal function, as the serum creatinine concentration is influenced by GFR and other "non-GFR determinants" including the muscle bulk, dietary intake, renal tubular secretion and extra-renal creatinine elimination by the gastrointestinal tract [18]. Creatinine based eGFR estimation has evolved from Cockcroft-Gault (CG) formula [19], Modification of Diet in Renal Disease (MDRD) [20], to CKD-EPI equation [17]. KDIGO 2012 guideline recommended the use of the CKD-EPI equation for evaluation of eGFR in adults [2]. Hospital Authority (data not shown in the results section). Defined by eGFR < 60ml/min/1.73m 2 , the prevalence of CKD varies throughout the world. In Europe, it varies between 1.7 to 11.5% in the general population, and 2.2-14.3% in hypertensive patients [13]. In the United States, USRDS reported the CKD prevalence of 6.9% in the general population and 16.1% in the hypertensive population among participants of the National Health and Nutrition Examination Survey (NHANES) 2013-2016 [14]. In Taiwan, a local cohort based study found 9.1% CKD prevalence in the general population and 26.0% in hypertensive individuals [15]. There were heterogenicity in patient source and sampling (electoral rolls, general practitioners lists, cohort, etc.), age ranges (all ages or elderly), eGFR calculation methods (CG, MDRD, CKD-EPI equations), test frequencies (single test or repeated tests), and definition of CKD (by eGFR calculation or by diagnostic codes). In our study, we found that CKD was present among 15.3% adult Chinese hypertensive patients, that was similar with those reported in the US and some European countries, but higher than other European countries and lower than Taiwan. The discrepancy could be due to the true difference or method diversity.
Although univariate analysis showed more possible associated factors could be related to CKD, the multivariate analysis after adjustment showed that significant factors associated with CKD were older age (OR 3.56 for every 10 years increase, p<0.001), history of CHF (OR 6.26, p 0.032), DM (OR 1.73, p 0.031), gout (OR 3.02, p 0.021), lower HDL (OR 0.29, p 0.001), and presence of proteinuria or albuminuria (OR 2.72, p <0.001). Firstly, our study showed a strong positive correlation between older age and increased risk of CKD, which is consistent with previous studies both in the general population [21] and in hypertensive patients [22][23][24]. There is a debate whether decreased GFR in older people represents an actual disease or a "normal ageing" phenomenon, as GFR declines steadily wit h ageing, beginning at age 30-40 years, with an apparent acceleration in the rate of decline after age 65-70 years [25].
Furthermore, glomerular sclerosis, tubular atrophy and vascular sclerosis are associated with ageing [26]. Given the fact that there appears to be increased risk of complications associated with decreased eGFR in older people irrespective of cause, KDIGO considers all individuals with persistently decreased GFR less than 60 ml/min/1.73m 2 to have CKD, which is still the current standard of practice and research. History of CHF was found to be a strong associated factor for CKD, a similar finding as supported by other studies [27,28]. Actually sometimes CHF and CKD are considered concurrent chronic disease epidemics [29]. CHF as the primary syndrome can experience secondary CKD, and vice versa, or both can coexist on the basis of shared risk factors. In this cross-sectional study, it is hard to tell which disease is primary and which is secondary. It was not unexpected that DM was an associated factor with CKD, as DM itself is the leading cause of CKD and ESRD in developed countries. As a well-recognized microvascular complication, kidney impairment develops in approximately 30% of Type 1 DM patients and 40% Type 2 DM patients. Among the risk factors for Diabetic kidney disease initiation and progression, hyperglycemia and hypertension are the two most prominent factors [30]. An association between gout and CKD has been recognized for many years [31][32][33]. The association could be bidirectional, with CKD as an independent risk factor for gout [34] and gout patients potentially predisposing to CKD possibly by hyperuricemia, chronic inflammation or NSAIDs drug therapy. Some interventional study suggested urate lowering treatment may have a beneficial role in renal function protection [35,36]. Dyslipidemia is common but not universal in CKD patients. The presence of dyslipidaemia was affected by eGFR, presence of DM, the severity of proteinuria and nutrition [37]. While KDIGO Work Group no longer recommended LDL-Cholesterol as the single indication or target for pharmacological therapy, some studies supported the role of low HDL-cholesterol in the development and progression of CKD [38,39]. However, the protective role of HDL in CKD is being challenged and needs further evidence [40]. It was not surprising to find proteinuria or albuminuia was a strong associated factor of CKD. Actually albuminuria has been recognized essential in cardiovascular risk stratification in CKD patients. Proteinuria or albuminuria is not only a marker of kidney injury, but also a potential toxic contributing to renal function decline [41]. KDIGO [2] suggests initial testing of proteinuria in the following descending order: 1) urine Albumin-to-Creatinine Ratio (ACR); 2) urine Protein-to-Creatinine Ratio (PCR); 3) reagent strip urinalysis for total protein with automated reading; 4) reagent strip urinalysis for total protein with manual reading. A spot urine protein sample for protein with standard urine Dipstick test was recommended by AFCKDI [12] as a convenient and cost-effective tool for proteinuria detection in the primary care setting.
Although not all patients were tested for albuminuria in our retrospective study, the data on presence of either proteinuria or albuminuria apparently implied the positive relationship between them and CKD.

Limitations of the study
There were limitations to this study. Firstly, not all HT patients were checked for urine albumin. Dipstick for urine protein was used instead, which could be less accurate, and limited further risk stratification according to ACR levels. Secondly, this was a single centre data from a public primary care clinic. Although the sex ratio and median age are similar among different urban districts in Hong Kong, and the ethnic distribution imbalance was offset by excluding non-Chinese population, selection bias could still exist. Larger scale multicenter study was expected to overcome this limitation. Thirdly, patients in more advanced CKD stages would be referred to secondary care, thus the percentage of CKD 4/5 patients could be underestimated. The results may not be applicable to the private sector or secondary care setting. Lastly, given the cross-sectional design of the study, it could not establish a causal relationship between associated factors and CKD development. Prospective cohort study or interventional study would help provide more information on this regard.

Conclusion
We conclude that in Chinese hypertensive patients followed-up in the public primary care clinic in Hong Kong, the prevalence of chronic kidney disease with eGFR being less than 60ml/min/1.73m 2 was 15.3% by CKD-EPI equation. The prevalence showed an apparently increasing trend in elderly age groups. The associated factors for CKD were older age, history of CHF, DM, gout, low HDL level, and presence of proteinuria or albuminuria. Since CKD is a well-established risk factor for several clinical outcomes, family physicians should enhance their awareness of the high prevalence of CKD among HT patients and pay particular attention to the presence of the above associated factors. A concerted effort should be made in early recognition of risky CKD group in HT patients.

Ethical approval
The study was approved by the Cluster Research Ethics Committee. Ref: KC/KE-18-0196/ER-1. This is an observational study collecting existing data via Clinical Management System Retrieving Software without sensitive or identifiable personal information (name or ID), without affecting patient's management, and reported in